ClawHub
Back to skill

Security audit

Composite Scene

Security checks for vulnerabilities and agentic risk

Overview

This skill is a narrow image-composition guide that tells the agent how to combine user-provided reference images through a disclosed image-generation workflow.

Before installing, be aware that using this skill may send the reference images or image URLs you provide to the configured image-generation service. Avoid using sensitive private images unless that provider and workspace policy are acceptable for your use case.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Unbounded Output

Medium
Category
Output Handling
Confidence
60% confidence
Finding

Output size or generation rate is not bounded. Unbounded output enables denial-of-service through resource exhaustion, log flooding, or context-window stuffing.

Content

Scanner excerpt · SKILL.md (reported line 14)May include surrounding context.

md
# Composite scene

Combine several real elements (a product, a subject, a backdrop, a style) into one coherent image, with lighting and perspective already reconciled, without cutting out, masking, or compositing by hand. Each element is a separate reference image and the prompt directs how the pieces fit. This is the reverse of `character-consistency`: instead of holding one subject across many images, it pulls many images into one.

## Inputs to collect

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest describes this skill as merging several real images into one coherent picture, with examples centered on placing products/subjects into scenes and explicitly routing single-image edits elsewhere. Recipe 3 instead documents repainting a photo in the style of a painting, which is a broader image transformation task rather than combining multiple real images into one composite scene.

Content

No source excerpt is available for this finding.